| Algae are the oldest plants and most algae live in water.Algae not only has the important significance to provide the nutrient for fishery,but also can provide the basis for judging whether the water is polluted by the number of algal cells in the water.In the study of algae recognition technology,the traditional research steps consist of microscopic observation,morphological analysis and counting statistics.Mainly through the human eye observation,easily lead to visual fatigue,low efficiency,and the data can not be saved.As the development of modern digital image processing technology,algae cells can be identified and analyzed.Compared with traditional optical holographic,digital holography has the advantages of low production cost,fast imaging speed,recording and reproduction flexibility,and its recording and reproduction process can be digitally processed.Hog feature extraction of holographic reproduction of algae samples was carried out.Combined with SVM supervised learning model,algae cells were efficiently and conveniently counted and counted.In this paper,the basic theory of digital holographic lensless imaging is analyzed.Aiming at the characteristics of algae cells,a simple and easy lensless holographic imaging device was fabricated based on the theory of lensless holography.The algae is counted and analyzed on the basis of this device.The details of this paper are as follows:1.In view of the shortcomings of the traditional algae detection,this paper uses digital image processing technology to study algae cells.2.Introduces the principle of lensless holographic imaging and shows the optical path diagram without lens holography.Based on the theory of lensless holography and the study of algae cells in this paper,a holographic impermeable imaging device which can be easily carried and light in size and small in size is designed and made by using 3D printer.Device.Introduce the specific operating procedures and compare the advantages of this device compared to conventional devices.3.In order to carry out rapid and efficient feature extraction on algal cells,algae were studied by HOG descriptor.Since HOG is manipulated on the local squares of the image,it maintains a good invariance of the geometrical and optical deformationsof the image.As long as the coarse air sampling,fine sampling of the direction and a strong localized optical normalization and other conditions,algae cell detection effect will not be affected by other factors.So the HOG feature is particularly suitable for doing algae detection in images.4.Using SVM to supervise the learning model,algal cells were identified and counted and analyzed.Compared with the traditional learning methods(such as pattern recognition,neural network),SVM is based on the principle of structural risk minimization.It is a convex optimization problem,so the local optimal solution must be the global optimal point.In addition,SVM solves the linear and non-linear classification problem. |
